Image Compression Using Adaptive Wavelets and Trellis-Coded Quantization
In this work we present some new results on still image compression. The first step involves adaptive transform (AWT) as the fundamental means of de-correlation of the image pixel values. The wavelet decomposition scheme is an adaptive one, where at each point a locally optimal decision is taken regarding whether a wavelet decomposition step is to be applied and also the direction of the filter, (along image lines or columns) and the choice of wavelet filter, using the optimal bit allocation formula. The same formula also determines the scanning oder and the bit rate allocated for each resulting wavelet subband. The next step is the application of multistage Trellis-Coded Quantization (MTCQ) on the various subband. To do so, we use a family of MTCQ quantizers, parametized by the target bit rate. Progressive transmission is achieved by quantization of the residual error.
KeywordsEntropy Autocorrelation Convolution Pyramid Univer
Unable to display preview. Download preview PDF.
- 2.I. Daubechies, Ten Lectures on Wavelets, Philadelphia, PA: SIAM, 1992.Google Scholar
- 4.A. Gersho and RM Gray, Vector Quantization and Signal Compression, Kluwer Academic Publishers, 1992 Google Scholar
- 5.YBHoung-Jyh Wang and C-C J Kuo, “Multi-Threshold Wavelet Coded (MTWC),” in JPEG2000 Document, WG1N665, November 1997 Google Scholar
- 6.S. Lin and DJ Costello, Error Control Coding: Fundamentals and Applications, Prentice Hall, 1983 Google Scholar
- 8.WB Pennebaker and JL Mitchell, JPEG - Still Image Data Compression Standard, New York: Van Nostrand Reinhold, 1993 Google Scholar
- 14.H-J Wang and C-C J Kuo, “A Multi-Threshold Wavelet Coder (MTWC) for High Fidelity Image,” in IEEE Signal Processing Society, 1997 International Conference on Image Processing (ICIP 97), July 1997 Google Scholar